JayWeintraub.com - Internet Advertising Analysis and Growth Insights

PriceGrabber (Up for Grabs III)

For those looking for additional insights on the PriceGrabber acquisition, I recommend David Lewis' prescient Shopping Comparison Scorecard post from June, the month when theShopping.com and Shopzilla acquisitions occurred.

Do no evil? Try, No mercy.

If you have ever seen a mortgage lead form, you will know that one of the drop downs deals with the potential borrower’s credit. Not being a mortgage banker or broker, it would seem reasonable to believe that those on the lending side would only want to deal with people who had good credit. - they will not have a problem getting the loan and will borrow more money. Both are true, but the average borrower with good credit also tends to shop around more. They have a greater understanding of their own value and of what their loan should look like. As a result, while often-good clients, they aren’t highly profitable ones. Borrowers with poor credit on the other hand, do not shop around as much and are more likely to take whatever deal they are given, and lenders have certainly taken advantage of this fact. When abused, this practice is known as predatory; on the whole, it’s only slightly deceptive. Whether good or bad, the truth is that companies do this all the time. One company that you might not expect to benefit from this, or facilitate it, is Google.

Most buyers and sellers have a familiarity with paid search. It involves the selection of keywords and entering a price for those keywords. Think of it like eBay. Each keyword is an item, and the process for having your ad associated with that keyword involves a dynamic auction. Unlike eBay where auctions involve only one item and thus one winner, in paid search, each keyword can have up to 40 advertisers associated with it. Those who bid more get listed higher, and the higher the position, the more traffic one receives. Yahoo’s Overture division popularized this process. Like eBay, they let advertisers bid for position and each bidder knew what the bidder in front of them and behind them paid. Pay more and you can get listed higher. As it’s an auction, once you raise your bid, the other person might too. You can’t know for sure what their limit is, but you will always know the price it takes to move up or down.

For those who have not spent or made money with Google, their system works differently. They too allow advertisers to bid on keywords, and placement involves a dynamic auction. The biggest difference though is that Google includes a performance criterion along with price when determining the ranking of the advertisers. The advertiser in the top position might pay more per click than the advertiser in the second position, but there is no guarantee. Paying more often helps to get listed higher, but just how much more is undeterminable. And, unlike Yahoo, Google does not disclose the price per click of the advertiser.

Google’s methodology for bidding is known as a sealed auction. Everyone bids, but what the winning bid is, is unknown. Yahoo on the other hand, by disclosing the bid prices, runs the opposite, an open auction. While simplistic, given the experience level of most readers, this context on search – from the basic description of pay per click and the various auction methods employed by the engines – explains the secret to Google’s profit. By not knowing what others bid, advertisers have no choice but to experiment. They bid on more words and at different price levels in hopes of getting the position they want. Experimentation is inefficient and as a result highly profitable. But, what about the publishers?  How does this impact them? By not revealing their bid amounts, publishers running Google ads on their site have no real way of determining what revenue share they receive. And, even if on a fixed revenue share, they still have no way to gain insight into the performance of the keywords that made them their revenue.

Google’s sealed auction method, one that combines price with performance data but shares neither has real benefits. It does promote better ads, a more active marketplace, and it gives the company the flexibility to display the ads across various types of content. They can adjust the prices charged to the advertiser based on their assessment of the traffic quality, and it also means they can pay publishers varying amounts for the same ads based on Google’s perception of the site’s quality. The downside though, and why Google isn’t all that different from consumer financial service companies that work with poor credit borrowers is that they don’t help anyone save money, or pass the extra money they receive directly on to the publishers.

The new Microsoft continues to roll out, as the market demands, making sure not to increase their closed wall nature. This past week, for example, Google announced free analytics for advertisers, the ability for advertisers to bid directly for placement on a particular site while visiting that site, and separate pricing for content. But, don’t be fooled. Google does what is good for Google. They will not make it any easier or cheaper for the marketers or any more profitable for the publishers than it has to be. Do no evil – perhaps to the end user, but for advertisers and publishers, it’s more like “No mercy.”

---
An "Aussie Webmaster" :) said the following in a comment on this article at DM Confidential, "Watch out when Yahoo adds the silent bid technology to its stuff early next year."

Creating the next Valueclick

Could a company that generates per year what a similar company does per month, one whose market cap equals that bigger company’s quarterly revenues, be considered the “next” iteration of it? That appears to be the question that CGI Holdings Corp., d/b/a Think Partnership might just be answering. At a run rate now surpassing $20 million per year and a current market cap of $90 million, Think Partnership still has a long way to go before they challenge the company they most closely emulate, Valueclick, but daunting a task as that might seem, especially for one that did not have the benefit of an IPO during the early internet heyday, the little company that could might just pull it off. This week, we’ll take a closer look at Think Partnership, a company that we wrote about back in March when they first appeared on the industry radar with their acquisition of web marketing company, Primary Ads.

Reading through their history, CGI Holding Corp doesn’t seem like a company that would be where it is today. They say this about themselves in a filing this year, “We have lost money historically. We had net losses for the years ended December 31, 2002 and 2001. Our future operations may not be profitable. If we are not profitable in the future, the value of our common stock may fall and we could have difficulty obtaining funds to continue our operations. Our balance sheet is weak. We lack the capital to compete aggressively. Our growth is capital constrained.” Granted, those statements are in the risk factors section, but they don’t give the immediate impression of a company that will challenge or be similar to Valueclick. But, those statements, while true, are also perhaps overly modest and cautionary. Think Partnership’s story is one of growth, leveraging a good market and knack for acquisitions.

CGI Holdings is not a new company. It was incorporated in 1987, but it didn’t really become active in the space until 1997 but more so in 2001 when it purchased Internet stalwart but outdated Worldmall.com. That was renamed Websourced in 2002 and became a division of CGI Holdings, focusing on search engine marketing. Websourced was best known by one of its subdivisions, KeywordRanking.com, run by Patrrick Martin, who founded Worldmall.com. In 2003, Think Partnership, under its Websourced division launched affiliate dating site Cherish.com. The two, along with the upturn in the Internet advertising economy helped the parent company to achieve strong revenue growth in 2004, this after becoming profitable in Q2 2003.

The real fun began in 2004. That June they made their first of many acquisitions to follow, buying Engine Studio, creators of a product to help small businesses get listed on the search engines. Engine Studio’s assets presumably helped beef up the competency of Websourced and the value perception of Think at a time when Google was getting ready to go public. And, thanks to the strong revenue growth coming from their search business, their only division at that time, CGI decided not to push ahead with its reverse stock split. They did continue buying, announcing the acquisition of WebCapades in August. Then, in November, they signed a letter of intent to merge with privately held MarketSmart, which ran three companies. That merger led to Think’s dramatic rise in 2004 revenues to $20 million, up from $7 million in 2003.

Just before the close of the year, Think outlined their future strategy, raising $15 million in a private placement offering to help continue their aggressive acquisition strategy. Shortly thereafter, also in December, the company declared its intent to merge with another, Proceed Interactive, at the very end of 2004. That deal ultimately fell through but the buying spree marched forward. In February of 2005, Think purchased two other companies – Personals Plus, which extended their dating business, and Ozona Online, a small web shop that does design, hosting, and custom builds. The price was $2 million and $300,000 respectively. Also in February, in a win for the group, its stock moved from being an over the counter to the American Stock Exchange, trading under THK.

Things got really interesting in March when they announced their desire to purchase affiliate marketing software company Kowabunga - a scaled down version of Valueclick’s Commission Junction. They followed that up in April with a merger with Vintacom and Real Estate School Online in deals worth almost $5 million each. Also in April, Think announced its desire to purchase Primary Ads for $10 million. That deal was followed up by its decision to purchase Babytobee.com. It was the Primary Ads deal that had most of us taking notice of Think, as their, now well documented but still under the radar, spat with Direct Technologies lasted several months this year. Think, though, marches strong, increasing their revenues and profits and now having a suite of skill sets from search engine marketing, to an affiliate platform, their own advertising campaigns, internal inventory on which to run them, and a company that can help promote theirs and other offers to third parties. No doubt, we will see more acquisition announcements from them in the future. Theirs is a trend not necessarily started by Valueclick, but they have thus far executed on an aggregation strategy that has taken a relative tiny player and turned them into a future force to be reckoned with. Others will no doubt follow, as this is just the beginning of the mass roll-up, and they a leading example.

Up for "Grabs" II

Price Grabber acquisition update - An earlier post mentioned Price Grabber being the next comparison shopping engine to be snapped up. That has been independently confirmed and the buyer is now known. While not able to write their name, I can say it is not Valueclick, MSN, News Corp., or Google/Yahoo. The buyer has substantial online assets and has been featured here before. Happy guessing.

Bankrate.com Buying Interest.com & Fastfind.com

Financial content bohemoth and web calculator tool provider Bankrate signed a definitive agreement to purchase two companies - Interest.com and Fastfind.com. Not being familiar with the former, Interest.com appears to be more mortgage-focused than Bankrate and a second-tier competitor to them. Their model relies more on site-based advertising rather than Bankrate's combination of site-based ads and licensing fees for use of their tools.

Fastfind.com, whose founders come from GetSmart (purchased by Lending Tree who bought it from Providian) makes sense for Bankrate but could worry companies like LoanWeb, LowerMyBills.com, NexTag, and other mortgage lead gen advertisers. While expensive, Bankrate's inventory provides quality leads. The acquisition of Fastfind gives Bankrate in-house competence to monetize their own traffic and take that margin as their own.  It looks like after years of seeing these advertiser make money off their site and now having the resources to do something about it, Bankrate has.

Bankrate has a market cap just north of $500 million and $40 million cash with no debt. Their revenues grew from $26 million in '02 to $36 million in '03 to $39 million in '04. The stock is up more than two-fold since April of this year. Analysts predict the company to do $57 million this year, and Bankrate has increased their guidance for 2006 to roughly $80 million.

Adteractive on the prowl

Adteractive has joined Quinstreet in acquiring companies that can help them own traffic. Adteractive recently acquired many education related sites created by Potera Media Group. Both companies are located in San Francisco and I'm guessing Potera was an affiliate of Adteractive, giving the latter good intel on performance and prospecting.   

Up for "Grabs"

As broken by Josh Stomel, Price Grabber will become the latest comparison shopping acquisition.

In June of this year, eBay bought Shopping.com for $620 million and EW Scripps bought Shopzilla (BizRate.com) for $525 million.

For a complete list of internet advertising deal flow, I recommend Cliff Kurtzman's Online Advertising Industry M&A Deal Flow.

The Ownership of Traffic

The first half of A Crash Course on the Digital Marketing Vendor Landscape, the basis to last week’s two part Digital Marketing 101, focused on the different types of internet advertising and the companies that operate in them. The second half of that presentation focused on something entirely different and something that  was not covered in last week’s introduction to digital marketing, and that is the numerous mergers and acquisitions. We have covered mergers and acquisitions, and the often corresponding consolidation many times, including most recently in Digital Thoughts - Following the Money, An Investment Recap - Part 1  and Part 2. Additionally, Cliff Kurtzman, with whom I gave the crash course presentation during Ad:Tech, does a phenomenal job of recounting the numerous deals here. This week’s Trends Reports though looks at the underlying reason behind many of those deals – the desire to own traffic, in particular the network / publisher consolidation.

Even the famed Intermix/MySpace.com acquisition had the net result of giving Rupert Murdoch’s News Corp. traffic. That certainly applies with the New York Time’s purchase of About.com, WebMD acquiring TheHeart.org, Yahoo buying Flickr, AOL purchasing Weblogs, eBay scooping up Skype, and the list goes on. In fact, of the tens of billions spent this year on mergers and acquisitions, half, if not more helped one company purchase the user base of another. As Cliff has pointed out, while $580 million might seem like a lot for a social networking site, a company could spend a lot more over a longer period of time and not build up an active user base of more than 20 million people who together produce a site with the most ad impressions.

The size of the MySpace deal and many of the others that involved the acquisition of a user base is distracting. It’s easy to focus on the hundreds of millions of dollars changing hands and not the underlying trend. Some in our space have taken notice and are the ones secretly leading the strategic consolidation trend, i.e. purchasing profitable companies that currently contribute to their bottom line. These deals tend to be less than $10 million and are all aimed at owning traffic. Thanks to 2004 and 2005, which treated most of the networks well, many have a nice war chest and can now consider corporate development activities. And, while not the reason for becoming a network, their business model has given them an advantage with respect to potential acquisitions. They connect publishers with ads. This means they not only know a lot of publishers but also know how much money that publisher can generate.

So far, two reasons seem to drive the networks that have started to acquire. The first has done so simply to lock up distribution and defer margin compression. When they started, most new ventures online focused on building traffic not monetizing it. Entrepreneurs made web sites, created email lists, and even distributed ad ware. Today, the gap no longer revolves around traffic but monetization. The barriers for creating a new network have decreased while the sophistication of the entrants has increased. In the current environment, we see those who once had, or still do have, the traffic turning into the aggregator, the network. Google is a prime example. Access to traffic is only getting tougher, i.e. expensive and margins thinner.

The second reason for owning traffic relates to margin compression insulation, it deals with reduction in potential volatility. Several companies that had a good ‘04 and ‘05 want to leverage the past success and overall promising Internet advertising market to have an initial public offering. And, most of these companies would define a successful offering by achieving a high and sustainable stock price. Two or three key publishers / affiliates moving could affect that price.  So, as insurance against such fluctuations, they have bought, through cash and stock, these traffic anchors.

In many ways the high level consolidation not only feels like, but also qualifies as, a bubble. Similar to the current offline, real estate bubble the right market conditions have led to inflated pricing. In both Bubble 1.0 and today’s Bubble 2.0 one of the main objectives in investing was to own traffic. The biggest difference between this bubble and the one that already burst is that many of these deals actually make sense. They might be overpaying for a Malibu beach home, but at least they’re not paying for swamp land. And this certainly holds true for the networks buying up their publishers. Bubble fever has driven up the cost of these publishers, but the logic for doing so not only makes sense, but also should prove a key factor in their longevity. So, while the answer to, if we are in a bubble is yes, at least we know the floor is more stable. The clever networks have joined in to help solidify theirs.

The Blurring of Search and Media

That Ad:Tech might have felt crowded, overlapping, and utterly confusing – reflects not on the show but the industry. Our industry used to have easily defined boundaries and companies that could fit into independent silos. Like the continents, the landmasses of our industry have shifted over time, blurring the lines between areas that once felt so separate. Of all areas, search and media best exemplify this blurring. This week we focus on how those two landmasses have shifted over time, and we do so by looking at these sectors from mid-2000 until June of 2003 when Google launched its AdSense program for publishers.

In 2000, media meant the banner and the pop-up. Skyscrapers, leaderboards, tiles, and the multitude of rich media units didn’t exist and/or hadn’t gained any traction. The advertiser mix was not incredibly diverse and serving costs too high to do anything outrageously interesting. Per impression and per click pricing accounted for a bulk of the impressions, with cost per action just starting to gain acceptance from the sites and networks that had the foresight to see the arbitrage opportunities opening up. This time was the beginning of a major change in the ad network space. Of the most popular in early 2000 -  L90, Engage, Flycast, Valueclick, and DoubleClick, only Valueclick exists today. Advertising.com, while in existence for two years, did not enjoy the visibility that it does today from advertisers and publishers.

The majority of the ad networks at that time made money by playing matchmaker between advertiser and publisher. Ad networks today still do this, but those today do three things fundamentally different. The first is that they assume more risk, this is to say that ad networks today will accept payment on one metric (ex: CPA) but pay publishers on another (ex: CPM). In 2000, only Advertising.com was actively starting to embrace this model. The second difference relates to the methodology that companies use for choosing what ads to serve where. Today, it is commonplace and almost expected that companies use sophisticated problem solving to determine scheduling of the ads. The assumption of risk and the need to extract maximum value from available ads drove this model.

Regarding the third distinction between today’s ad networks, many original ad networks wanted to play more than a supplemental role. They wanted to act as the agent of record for a websites inventory, some choosing to have long contracts locking up the industry. As supply no longer became the constraint, those companies who positioned themselves as gatekeeper, had no model. They did not invest in the necessary technology or prepare for a role as anything but sole provider. The websites who had initially embraced an outsourced sales force quickly became disenchanted upon seeing their flow of ads dry up. Advertising.com, who never had a site representation approach, and along with Valueclick did not go under when the higher priced impression ad dollars left. Tribal Fusion and Burst Media, two other site representation companies and ones that began during the earlier years of Internet advertising, did survive though. Today’s more prominent ad networks, including Fastclick (now part of Valueclick Media), Casale, and Revenue.net, all began during the down times.

Media, until the middle of 2003 and even extending into the beginning of this year, has carried with it an association with graphical ads of set sizes. Today, while many sites still adhere to these sizes, graphical ads no longer account for all of the impressions. On many sites, especially content ones, text ads within that space are just as likely to appear if not more so than a graphical one. More than that, the text ads that appear are not driven by the typical “media” companies (i.e. the ad networks) but the “search” companies like Google and Yahoo, most notably. While more than Yahoo (then Overture) existed in 2000, no other company came close with respect to the number of advertisers, amount of keywords being bid on, and the total volume of clicks.

Having built up an immense following from their better search, Google had the platform by which to launch the first serious competitor to Overture, which it did in October 2002. The convergence of search and media began shortly thereafter. What Google executed in scale before others was treating media as a natural extension of its current advertising business. Unlike companies that focused on graphical ads who typically serviced a few hundred clients, Google serviced tens of thousands. They might not have known who each advertiser wanted to target, but they knew what they wanted. They knew the context, i.e. the keywords.

Rather than thinking of keyword targeting ads as only applicable to search, an activity that passively relies on users to enter queries, they began to look for proactive ways to assess user interest. In part due to their acquisition of Applied Semantics, Google developed the ability to map keywords to content. Their contextual technology allowed them to scan the content of a page, determine its meaning, and find keywords that related to the meaning. Those keywords just so happened to already have paying advertisers for them too. Voila, more places to show their existing advertisers.

Google wasn’t the first to try to put search (context / keyword) ads on web sites. The problem with earlier attempts was that the matching did not generally correlate to the content. High priced words were placed on any content regardless of relevancy, which quickly drove down the bid prices thanks to the market dynamics that search companies created.  Even when relevant, content placements do act differently than search placements, something Google figured out and created a solution for in their Smart Pricing system that lowers the actual bid values charged to advertisers based on where the ad runs.

Traditional media companies, those focused on graphical ads, often offer and have offered channel level targeting, but their models cannot support the granularity of targeting offered by contextual advertising. And that gap that began in 2003 most likely won’t close any time soon. That doesn’t imply a full convergence of search and media. Each has its strengths. What Google did was simply realize that the gap between the two didn’t have to be as great and that they could offer one interface for advertisers to reach their users. In doing so, they have changed expectations of ads so that no longer does a 468x60 space imply that a graphical ad will show. Google didn’t cause search and media to converge, but they have sped it up as they attempt to create their own Pangea of all forms advertising. And this is exactly what they will do and what we are seeing with other forms of Internet advertising that typically seemed like two separate continents.

NY Times Article: Google Wants to Dominate Madison Avenue, Too

From the NYTimes.com - an article on my Reading List that I post here as it is no longer readily accessible.

October 30, 2005

Google Wants to Dominate Madison Avenue, Too

 

By SAUL HANSELL

   

Mountain View, Calif.

IN many ways, Larry Page and Sergey Brin seem an unlikely pair to lead an advertising revolution. As Stanford graduate students sketching out the idea that became Google, the two software engineers sniffed in an academic paper that "advertising-funded search engines will inherently be biased toward the advertisers and away from the needs of consumers."

They softened that line a bit by the time they got around to pitching their business to venture capitalists, allowing that selling ads would be a handy safety net if their other, less distasteful ideas for generating revenue didn't pan out.

Google soared in popularity in its first years but had no meaningful revenue until the founders reluctantly fell on that safety net and started selling ads. Even then, they approached advertising with the mind-set of engineers: Ads would look more like fortune cookies than anything Madison Avenue would come up with.

As it turned out, the safety net was a trampoline. Those little ads - 12 word snippets of text, linked to topics that users are actually interested in - have turned Google into one of the biggest advertising vehicles the world has ever seen. This year, Google will sell $6.1 billion in ads, nearly double what it sold last year, according to Anthony Noto, an analyst at Goldman Sachs. That is more advertising than is sold by any newspaper chain, magazine publisher or television network. By next year, Mr. Noto said, he expects Google to have advertising revenue of $9.5 billion. That would place it fourth among American media companies in total ad sales after Viacom, the  News Corporation and the  Walt Disney Company, but ahead of giants including NBC Universal and Time Warner.

Not content to just suck advertising dollars from Web search, Google is using its windfall to pay for an eclectic range of ambitious projects that have the potential to radically disrupt other industries. Among other things, it is offering to build a free wireless Internet network in San Francisco, plans to scan nearly every book published and is testing a free classified advertising system it calls Google Base.

More quietly, Google is also preparing to disrupt the advertising business itself, by replacing creative salesmanship with cold number-crunching. Its premise so far is that advertising is most effective when seen only by people who are interested in what's for sale, based on what they are searching for or reading about on the Web. Because Google's ad-buying clients pay for ads only when users click on them, they can precisely measure their effectiveness - and are willing to pay more for ads that really sell their products.

HIDDEN behind its simple white pages, Google has already created what it says is one of the most sophisticated artificial intelligence systems ever built. In a fraction of a second, it can evaluate millions of variables about its users and advertisers, correlate them with its potential database of billions of ads and deliver the message to which each user is most likely to respond.

Because of this technology, users click ads 50 percent to 100 percent more often on Google than they do on  Yahoo, Mr. Noto estimates, and that is a powerful driver of Google's growth and profits. "Because the ads are more relevant," he said, "they create a better return for advertisers, which causes them to spend more money, which gives Google better margins." (Yahoo is working on its own technology to narrow that gap.)

Google already sells its text ads for many other sites on the Internet (including nytimes.com), and is also moving tentatively to sell the picture-based interactive advertising preferred by marketers who want to promote brands rather than immediately sell products. Now it is preparing to extend its technology to nearly every other medium, most significantly television. It is looking toward a world of digital cable boxes and Internet-delivered television that will allow it to show commercials tailored for each viewer, as it does now for each Web page it displays.

Eric E. Schmidt, Google's chief executive, explains the company's astounding success in advertising - and reconciles it with the founders' distrust of hucksterism - by suggesting that advertising should be interesting, relevant and useful to users. "Improving ad quality improves Google's revenue," he said in an interview at the company's headquarters, known as the Googleplex. "If we target the right ad to the right person at the right time and they click it, we win."

This proposition, he continued, is applicable to other media. "If we can figure out a way to improve the quality of ads on television with ads that have real value for end-users, we should do it," he said. While he is watching television, for example, "Why do I see women's clothing ads?" he said. "Why don't I see just men's clothing ads?"

The media and advertising industries certainly see a future in which television ads are aimed at individual viewers. But few outside of the engineering Ph.D.'s at Google think that television ads should simply be utilitarian, rather than entertaining, provocative or annoyingly repetitive - the models that have worked so far. And some media industry executives wonder whether Google, which has already become the most powerful force in Internet advertising, should also become the clearinghouse for ads of all types - a kind of advertising Nasdaq.

"For all of us to throw all our eggs in the Google basket is dangerous, because no one should have that much power," said Jeff Jarvis, a veteran magazine editor who publishes BuzzMachine, a blog about the media, and is a consultant for About.com, a division of The New York Times Company. He added that if Google were to expand its ad sales to other media outlets, prices would fall. "Google commoditizes everything," he said.

There is no better example of that than Google Base, a service that allows users to post all sorts of information free, including classified ads, he said. Newspapers, which increasingly use Google to sell ads on their own Web pages, will see Google Base as a "frontal assault" on their lucrative classified-ad business, and they will say, "I can't trust Google," Mr. Jarvis said.

Mr. Brin said that preliminary versions of Google Base leaked onto the Internet and that the company's partners should not fear it. "Google Base is as much about classified as it is about zoology," he said.

Larry Page and Sergey Brin were exceedingly ambitious from the day they started Google, but the job of finding some source of revenue fell to Omid Kordestani, an amiable former Netscape sales executive who was brought to the company in 1999 by K. Ram Shriram, another Netscape alumnus and an early Google investor. Mr. Kordestani explored a range of ideas, including charging users for searches as well as selling Google's technology to corporations or to other Web sites - notably Yahoo - that were less shy about selling ads.

Eventually, in 2000, Google started to sell ads on its own site, but they were only a few lines of text placed above the search results. There were no graphics and no banners. At first, these ads - and later, a second form of text advertisement that ran down the right side of the page - were sold at fixed prices. But such an approach would not last long.

In early 2002, a Google employee, Salar Kamangar, now 28, convinced Mr. Schmidt and the founders to switch to an auction-based system like the one set up by Bill Gross, the head of IdeaLab. Mr. Gross had created Goto.com, a search engine made up entirely of ads, where advertisers paid only if their ad was clicked on, and the advertiser who bid the most per click was listed first. (Goto was later renamed Overture Services and then bought by Yahoo, an early Google backer that has become its fiercest rival.)

Mr. Kamangar, though, had an important improvement on the model. Rather than giving priority to the advertisers that bid the most per click, as Goto did, he realized that it was better to save the front of the line for ads that brought in the most money - a combination of the bid and the number of clicks on the ad. This was not only more profitable, but it also linked readers to ads that were more relevant to them. He also figured out that the system should use what is called a Vickrey auction - that is, to charge the winner only one cent more than the second-highest bidder. That gives advertisers an incentive to bid high, knowing that they will not be penalized if they are far higher than the rest of the market.

Mr. Page and Mr. Brin were suspicious of any system that put high-bidding advertisers at the top, Mr. Kamangar said. "They thought if someone was willing to pay more it was a negative," he recalled. But he was able to convince them that the site could be improved by incorporating how often users clicked on an ad.

Mr. Schmidt, who was still new as chief executive, was worried more that moving to an entirely auction-based system - amid a recession in online advertising - could be financially disastrous. "I said to Salar, 'Promise me the revenue won't go down,' " Mr. Schmidt said. "I was afraid people would realize these ads were worthless." In fact, revenue quickly increased tenfold.

As Google's audience took off, advertisers came running - many thousands of smaller ones at first, but soon large companies as well. Among Google's largest advertisers is eBay, which has long bought keywords for nearly every sort of merchandise it sells.

"The smartest thing that Google did was getting smaller advertisers to buy in," said Ellen Siminoff, the chief executive of Efficient Frontier, an agency that helps advertisers manage their campaigns on search engines. She estimates that Google has two to three times as many advertisers as Yahoo does, largely because Yahoo has a 10-cent minimum bid. This lets Google earn money on more obscure search terms for which rivals have no ads.

This growing advertising business gave Google the confidence to expand its audience. Most significantly, in 2002, America Online brought in Google to replace Overture, which provided both search and search ads; that deal enshrined Google as the premier search engine and ad network. Google won the deal by guaranteeing AOL a substantial sum, which it would not disclose. Google was willing to make that bid only because of its confidence in its advertising sales prowess. "If we were wrong," Mr. Kordestani said, "there were some scenarios that would bankrupt the company."

But by that point, Google had figured out that the same sort of computing and engineering skill that it used to find Web pages could also be used to improve the quality and, ultimately, the profitability of advertising. "Initially, we didn't understand how fundamental the computer science was in advertising," Mr. Schmidt said. "We didn't have enough staffing or focus on this area. I managed to fix that."

GOOGLE introduced its current system for determining which ad to show on which page late last year. It is a wonder of technology that rivals its search engine in complexity. For every page that Google shows, more than 100 computers evaluate more than a million variables to choose the advertisements in its database to display - and they do it in milliseconds. The computers look at the amount bid and the budget of the advertiser, but they also consider the user - such as his or her location, which they try to infer by analyzing the user's Internet connections - as well as the time of day and myriad other factors Google has tracked and analyzed from its experience with advertisements.

"If someone is coming from a particular location, a certain ad may be more popular there," explained Jeff Huber, Google's vice president for engineering. "The system can use all the signals available, and the system itself learns the correlations between them."

This technology is both amazing and potentially frightening. Google already collects and keeps vast amount of data about what Web pages and advertisements each of its users click on, and it can evaluate that history - and compare it with that of hundreds of millions of other users - to select the ad shown on each page. For now, Google says it identifies users only by a number in a cookie it places on each computer that uses Google. It says it has not connected the vast dossier of interests and behavior to specific users by name. But that could change as Google offers more personal services - like e-mail messages and social networking - and works more tightly with partners who already have such personal information.

Lauren Weinstein, the founder of the Privacy Forum, said the data that Google collects creates troubling privacy issues, especially because it declines to say what data it keeps or for how long. "If you start to target people based on a corpus of data, it can be abused in various ways internally and externally by organizations and government agencies," he said. Government investigators and lawyers in civil suits regularly get court orders to force Internet companies to reveal e-mail messages and other personal information about users.

Google recently rewrote its privacy policy to make it easier to understand what data it collects, but it did not scale back its data retention. Nor did it, as Mr. Weinstein and others have demanded, give users the right to see the data collected about them and their computers.

For now, the only personal information Google says it considers is the user's location, which allows it to display ads for local merchants. It is starting to encourage other Web sites to send it the ZIP codes of their registered users so Google can display ads relevant to their location.

Mr. Brin said he was not sure what other information about users might prove useful, but he said Google would not use the data inappropriately. "I don't think it's a big deal to show opera glasses to someone searching for binoculars that you somehow infer is a woman," he said. "But you don't want to pop up ads for H.I.V. drugs on someone's page, because you inferred they have H.I.V., when their boss is standing there looking at their computer."

To be sure, other Web sites are far more aggressive in using personal information. Yahoo will let marketers display ads to users based on demographic information the users provide as well as the users' surfing and searching history. Microsoft's new system for MSN explicitly allows advertisers to bid different prices for clicks from users of different ages, sexes and locations.

In addition to selling ads on its own site and on other sites that use its search technology, Google also places text ads on all manner of sites published both by professional media companies and by amateurs. Mr. Brin created this program in early 2003 after he became worried that the Internet crash would keep people from creating interesting Web pages for Google to index. This technology, called AdSense for Content, has made advertising on Google more attractive and provided the economic foundation for the rise of blogs.

"God bless Google," said Mr. Jarvis, the BuzzMachine blogger. "They took the cooties off citizen media." Until Google's program came along, advertisers shied away from placing ads on individual user's pages. But AdSense analyzed each page and tried - not always successfully - to find ads related to the page's content.

Now Google is looking to expand its advertising into even more places. It is testing a plan to buy pages in magazines on which to place text ads. And it also shows ads as users browse its new book search service. "A lot of the world's content is not accessible today and thus it is not easily monetizable today," Mr. Kordestani said. "We will figure out how to get more and more content and find the right way to put ads on it."

Advertisers, meanwhile, have had to scramble to adapt to this completely different approach to buying ads. They needed to find ways to keep track of bids on thousands of keywords, and to measure which ads, tied to which keywords, produced which sales - and then to figure out if they had bid the right amount for the ad.

Many advertisers and their agencies have a powerful love-hate relationship with Google. They find it a meaningful source of leads and sales, and the effectiveness of Google's ads is much easier to measure than that of traditional media. But Google has sometimes been hard to deal with. There is a growing sense that a significant number of clicks that advertisers pay for are fraudulent - made by competitors trying to deplete advertising budgets or by Web sites trying to bolster the revenue they get for displaying the ads. Google says it has technology to minimize what is called click fraud, but many people in the Internet business are skeptical that the incidence of fraud is as low as Google contends.

Here, as in other places, many advertisers criticize Google for being like a black box, because the company gives them less specific information and control than they would like. Until recently, for example, advertisers could not specify where their ads ran, though they were convinced that some Web sites in Google's network were much more likely than others to send them customers. Google responded with what it calls "smart pricing" technology that discounts certain ads if Google's analysis shows that they are seen on sites it determines are less likely to produce paying customers. But Google discloses little about how this works, and advertisers find it frustrating.

"Google is very opaque and bizarre to deal with," said Joshua Stylman, a managing partner at Reprise Media, a search advertising agency, but he added that Google had become somewhat more responsive in recent months.

Mr. Schmidt addresses those complaints by saying that advertisers are missing the point of Google's new model. It shouldn't matter what Google does with their ads, he argues, so long as the received value, which advertisers can measure, is higher than the price they pay. The entire discipline of media planning, which has long been important on Madison Avenue, may be rendered obsolete - just as Google's fully automated news Web site threatens the livelihoods of human news editors.

In any case, there is little doubt that Mr. Schmidt believes that science will replace much of the art of marketing. "I have this fantasy that goes like this," he said at one point. "You are the C.E.O. of a large company, and I come to you and say, 'Give me $1 million and give me your Web site, and we will guarantee you will get $100 million in sales.' Which C.E.O. would turn that down?"

Google isn't quite pursuing that sort of deal, but it is trying to have big retailers link their inventory systems directly to its advertising auction. That way, a toy store chain, for example, could respond to a search for dolls with an ad for either Barbies or Bratz, depending on which were overstocked in the store near the user's home. "Most retailers only advertise 5 percent of their products," said Tim Armstrong, Google's vice president for ad sales. "We can let them advertise all of them."

ON the other end of the spectrum, Google is also trying to focus on what the Internet market calls branding advertising - the sort that dominates television and magazines and creates awareness of a product, but doesn't directly call on viewers to buy right away. Yahoo, AOL and MSN have all evolved the simple rectangular banner ad into much more elaborate units with animation, interactivity and sometimes video formats that have been embraced by national advertisers.

Google has been able to convince some companies that its text ads can help build awareness of their products, even if people don't click on them to buy something. But top executives are also meeting weekly to develop a broader strategy for branding advertisements. Google has already allowed its so-called publisher network - those non-Google sites for which it sells ads - to accept advertising with limited graphics. At first, these were simple images, perhaps with a little animation. It is now moving to accept ads that use the popular Flash technology that allows for more interactivity. So far, these nontext ads have been only a tiny part of Google's business.

Indeed, such ads shine a spotlight on the mental compromise that Mr. Brin and Mr. Page made when they overcame their initial objections to advertising on their service. Text ads, they argued, were not the normal fluff of Madison Avenue, but actual information that was useful to searchers.

"Advertising was not a business built by logic, and we don't work by algorithm," said Wenda Harris Millard, Yahoo's chief sales officer. "Yes, we need to be more accountable, but that doesn't mean you sacrifice art and creativity."

Mr. Schmidt acknowledges that as Google explores moving into television, it may well face a conflict between its core belief that advertising must be useful and the typical television commercial that is "based on feeling and emotion."

"Our model is likely to affect television last," he said, while expressing optimism that a formula for useful, targeted commercials could be found. For now, he quickly added, the market for various forms of direct marketing is three times larger than that for television ads. "I was shocked by this," he said. "All of us are so conditioned to television as the height of advertising.

"We are in the really boring part of the business," Mr. Schmidt concluded, "the boring big business."